Abstract:The massive growth of network traffic data leads to a large volume of datasets. Labeling these datasets for identifying intrusion attacks is very laborious and error-prone. Furthermore, network traffic data have complex time-varying non-linear relationships. The existing state-of-the-art intrusion detection solutions use a combination of various supervised approaches along with fused features subsets based on correlations in traffic data. These solutions often require high computational cost, manual support in fine-tuning intrusion detection models, and labeling of data that limit real-time processing of network traffic. Unsupervised solutions do reduce computational complexities and manual support for labeling data but current unsupervised solutions do not consider spatio-temporal correlations in traffic data. To address this, we propose a unified Autoencoder based on combining multi-scale convolutional neural network and long short-term memory (MSCNN-LSTM-AE) for anomaly detection in network traffic. The model first employs Multiscale Convolutional Neural Network Autoencoder (MSCNN-AE) to analyze the spatial features of the dataset, and then latent space features learned from MSCNN-AE employs Long Short-Term Memory (LSTM) based Autoencoder Network to process the temporal features. Our model further employs two Isolation Forest algorithms as error correction mechanisms to detect false positives and false negatives to improve detection accuracy. %Additionally, covariance matrices forms a Riemannian manifold that is naturally embedded with distance metrices that facilitates descriminative patterns for detecting malicious network traffic. We evaluated our model NSL-KDD, UNSW-NB15, and CICDDoS2019 dataset and showed our proposed method significantly outperforms the conventional unsupervised methods and other existing studies on the dataset.

Zero-Day Traffic Identification Using One-Dimension Convolutional Neural Networks and Auto Encoder Machine

Abnormal Network Traffic Detection Algorithm Based on Improved Convolutional Neural Network

Detection of traffic congestion in road-occupied electric power construction based on video recognition

Deep Autoencoder Neural Networks for Short-Term Traffic Congestion Prediction of Transportation Networks

Real Network Traffic Collection and Deep Learning for Mobile App Identification

Elephant Flows Detection Using Deep Neural Network, Convolutional Neural Network, Long Short Term Memory and Autoencoder

A Novel Self-Supervised Framework Based on Masked Autoencoder for Traffic Classification

A Lightweight Intelligent Network Intrusion Detection System Using One-Class Autoencoder and Ensemble Learning for IoT

An Attention-Based Multi-Context Convolutional Encoder-Decoder Neural Network for Work Zone Traffic Impact Prediction

THREE-DIMENSIONAL CONVOLUTIONAL NEURAL NETWORK BASED TRAFFIC CLASSIFICATION FOR WIRELESS COMMUNICATIONS

Autoencoder-based Unsupervised Intrusion Detection using Multi-Scale Convolutional Recurrent Networks

Traffic-Sign Detection and Classification in the Wild

Hybrid Intrusion Detection Model Based on a Designed Autoencoder.

DxNAT - Deep Neural Networks for Explaining Non-Recurring Traffic Congestion

Industrial Control Malicious Traffic Anomaly Detection System Based on Deep Autoencoder

DarknetSec: A novel self-attentive deep learning method for darknet traffic classification and application identification

End-to-end encrypted traffic classification with one-dimensional convolution neural networks

An intrusion detection model to detect zero-day attacks in unseen data using machine learning

A Network Traffic Intrusion Detection Method for Industrial Control Systems Based on Deep Learning

Zero Day Threat Detection Using Metric Learning Autoencoders

An Encrypted Malicious Traffic Detection System Based on Neural Network